Difference between revisions of "Drum"
(Need optimum programming if used for main memory) |
(Tweak order, add LGP-30) |
||
| Line 1: | Line 1: | ||
| − | '''Drums''' were the predecessor [[magnetic storage]] technology to [[disk]]s; they were (as the name suggests) in physical form a drum | + | '''Drums''' were the predecessor [[magnetic storage]] technology to [[disk]]s; they were (as the name suggests) in physical form a drum. |
| − | + | They generally had a [[head]] per [[track]], like [[fixed-head disk]]s. Like them, on drums this avoided the delays involved in having the head(s) [[seek]] to the correct track, and was also mechanically simpler (an important point at this early stage of development), but did require more components. Because of their fundamental geometry, drums also needed more volume for their installation than disks did. | |
| − | The very first drums were actually used for [[main memory]] on some very early computers; and also on slightly later, but low-end, ones ( | + | The very first drums were actually used for [[main memory]] on some very early computers; and also on slightly later, but low-end, ones (notably the [[IBM 650]], [[Bendix G-15]], and [[LGP-30]]). Such use often required the use of [[optimum programming]], to control the execution delays (due to rotational latency) which resulted if [[instruction]]s were not placed in the optimum positions on the drum. Some drums were actually used to store [[register]]s, storing several registers in each track, and using multiple heads on that track to give faster access to the contents of the registers. |
Once [[core memory]] appeared, drums were only seen in the role of high-speed [[secondary storage]]; often being used for [[swapping]] or [[virtual memory|paging]] purposes. | Once [[core memory]] appeared, drums were only seen in the role of high-speed [[secondary storage]]; often being used for [[swapping]] or [[virtual memory|paging]] purposes. | ||
| − | Drums fell out of use because they took more physical room than a disk with the equivalent surface area. A number of [[disk platter]]s, giving in total a significantly greater amount of storage, could be stacked in the volume occupied by a single drum. | + | Drums fell out of use in large part because they took more physical room than a disk with the equivalent surface area. A number of [[disk platter]]s, giving in total a significantly greater amount of storage, could be stacked in the volume occupied by a single drum. |
[[Category: Device Basics]] | [[Category: Device Basics]] | ||
Latest revision as of 20:45, 18 March 2024
Drums were the predecessor magnetic storage technology to disks; they were (as the name suggests) in physical form a drum.
They generally had a head per track, like fixed-head disks. Like them, on drums this avoided the delays involved in having the head(s) seek to the correct track, and was also mechanically simpler (an important point at this early stage of development), but did require more components. Because of their fundamental geometry, drums also needed more volume for their installation than disks did.
The very first drums were actually used for main memory on some very early computers; and also on slightly later, but low-end, ones (notably the IBM 650, Bendix G-15, and LGP-30). Such use often required the use of optimum programming, to control the execution delays (due to rotational latency) which resulted if instructions were not placed in the optimum positions on the drum. Some drums were actually used to store registers, storing several registers in each track, and using multiple heads on that track to give faster access to the contents of the registers.
Once core memory appeared, drums were only seen in the role of high-speed secondary storage; often being used for swapping or paging purposes.
Drums fell out of use in large part because they took more physical room than a disk with the equivalent surface area. A number of disk platters, giving in total a significantly greater amount of storage, could be stacked in the volume occupied by a single drum.
